摘要
为应对化纤企业工艺生产过程和日常仓储中重大危险源过于集中而导致多米诺事故频发的现状,研究多米诺事故中不同灾种耦合效应对事故扩展的影响,采用贝叶斯网络模型对多米诺事故风险评估方法进行研究.首先,提出了多米诺事故耦合效应的定义,并对多灾种多因素耦合效应进行情景分析;其次,选用贝叶斯网络模型并结合Probit模型构建多米诺事故场景概率模型,从而确定了多米诺效应的演变过程、时间序列以及各层级事故单元间的影响作用,提高了多米诺事故风险评估结果的可靠性;最后,以某化纤企业储罐区为例,通过模型计算探讨了某储罐发生池火灾或蒸气云爆炸事故时的多米诺效应场景及最大可能传播路径,确定了多米诺事故扩展的可能性大小及关键单元,并在此基础上提出相应的控制措施.通过实例计算验证了该方法的工程可行性,结果表明,该方法能够在多米诺事故的预防与控制方面起到重要的指导作用.
In order to deal with the frequent domino accidents caused by the concentration of major hazard sources in chemical fiber enterprise’s process and daily storage, the influence of coupling effects of different types of disasters on the accident expansion was studied. Bayesian network model was used to study the domino accident risk assessment method. Firstly, the definition of domino accident coupling effect is put forward, and scenario analysis of multi-disaster and multi-factor coupling effect is carried out.Secondly, Bayesian network model and Probit model are used to construct a domino accident scenario probability model, thus determining the domino effect evolution process, time series and the influence among accident units at various levels, thus improving the reliability of domino accident risk assessment results. Finally, taking the storage tank area of a chemical fiber enterprise as an example, the domino effect scene and the maximum possible propagation path of a storage tank fire or vapor cloud explosion accident are discussed through model calculation, the possibility of domino accident expansion and key units are determined, and the corresponding control measures are put forward based on this. The engineering feasibility of this method is verified by an example calculation. The results show that this method can play an important guiding role in the prevention and control of domino accidents.
作者
张苗
宋文华
Zhang Miao;Song Wenhua(College of Textile Science and Engineering,Tianjin Polytechnic University,Tianjin 300387,China;Tianjin General Fire Brigade,Tianjin 300090,China)
出处
《南开大学学报（自然科学版）》
CAS
CSCD
北大核心
2019年第1期88-96,共9页
Acta Scientiarum Naturalium Universitatis Nankaiensis
基金
天津市科技计划项目(16ZXHLSF00290)
关键词
化纤企业
多米诺事故
耦合效应
贝叶斯网络
风险评估
chemical fiber production enterprise
domino accidents
coupling effect
Bayesian network
risk analysis
